Defining Cyclic Boundaries

[roughhawkbit]

There's something a little strange going on in the definitions of cyclic boundaries... as the number of grid points in any direction must be 2^n +1 I've always set the pointIn to be the same number and this is the case in the example protocol file single_species_single_substrate_2D.xml:

        <boundaryCondition class="BoundaryCyclic" name="x0z">
            <shape class="Planar">
                <param name="pointIn" x="0" y="-1" z="0"/>
                <param name="vectorOut" x="0" y="-1" z="0"/>
            </shape>
            <shape class="Planar">
                <param name="pointIn" x="0" y="33" z="0"/>
                <param name="vectorOut" x="0" y="1" z="0"/>
            </shape>
        </boundaryCondition>

However, this leads to agents seemingly off the agent grid... in this case the resolution is 8um and we have many cells with positions between 256um and 264um. Changing the "33" to "32" seems to fix this, but it may be causing other problems that I don't know about.

I'll try to get to the bottom of this eventually, but any help would be appreciated